From owner-chemistry@ccl.net Mon Sep 24 10:19:01 2012 From: "Colin Groom edir:-:ccdc.cam.ac.uk" To: CCL Subject: CCL: Statistical study of weak bonds in chemical compounds Message-Id: <-47653-120924094237-21500-LdzYUJFt10z+zWplyuHsJg---server.ccl.net> X-Original-From: "Colin Groom" Date: Mon, 24 Sep 2012 09:42:36 -0400 Sent to CCL by: "Colin Groom" [edir*ccdc.cam.ac.uk] Steve makes an important point, but despite the fact that hydrogen atoms do diffract X-rays weakly, the quality of crystal structures nowadays (using the latest X-ray sources, CCD detectors and processing software) is such that hydrogen atoms can be located quite accurately in a standard small organic crystal structure determination. The Cambridge Structural Database can be searched, via ConQuest, to retrieve just the type of information that you are interested in, Alexandre. The fragment is sketched and the geometric parameters required (for example, a distance from X to the amine N, the X-HN angle) are defined. The search will return the structures containing the fragment along with the values of the geometric parameters - which can be further analysed within a spreadsheet package or the CCDCs Mercury. There are at least 5000 hits for a substructure search of the type X-H----NH2. Whilst the vast majority of these contributing structures will be determined from X-ray there will be enough good quality information present in the dataset to allow you to probe the interactions you are interested in. You can use the X-N distance for example if you dont wish to include the H atom explicitly. If you need more help, do drop us an email at support^ccdc.cam.ac.uk we will be able to provide query files and give you more details on the set up of the searches. Colin Groom Executive Director Cambridge Crystallographic Data Centre Sent to CCL by: Steve Williams [willsd%%appstate.edu] CSD query may not give useful results even if it is well formulated. Nearly all of the structures in the CSD are X-ray diffraction structures. Since X-ray scattering power increases with electron density, and since the electron density of H atoms must be very low, X-ray methods are rather poor at locating H atom nuclear positions. There MAY be a few neutron diffraction structures in the CSD that will have the x-h-n motif that you seek, but if so there may not be very many of these. Steve Williams Sent from my iPad On Sep 22, 2012, at 5:15 AM, "Alexandre Borrel alexandre.borrel*_*univ-paris-diderot.fr" wrote: > > Sent to CCL by: "Alexandre Borrel" [alexandre.borrel _ univ-paris-diderot.fr] > Dear Members, > > We would like to do a statistical study of weak bonds in chemical compounds. > We are looking for data lengths and angles of hydrogen bonds X-H...N in first, secondary and tertiary amines. > 1. Can you suggest references of publications mentioning that ? > 2. Have you an idea of queries to retrieve in the Cambridge CSD bank the compounds with such bonding and locating the corresponding X,H,N atoms in the connection tables ? > > Thank you very much. > > Best regards, > > Alexandre Borrel. > > ----------------------------------------------- > BORREL Alexandre > PhD student University of Paris Diderot> > From owner-chemistry@ccl.net Mon Sep 24 10:58:01 2012 From: "G. Gopakumar gopakumar.gopinadhan+/-gmail.com" To: CCL Subject: CCL: Responses:- Dihedral angle: DFT vs Crystal structure Message-Id: <-47654-120924105348-25511-JnHS2dVHOQdm96apZsnjqQ__server.ccl.net> X-Original-From: "G. Gopakumar" Content-Type: multipart/alternative; boundary=20cf307f3996212c6a04ca73c118 Date: Mon, 24 Sep 2012 16:52:56 +0200 MIME-Version: 1.0 Sent to CCL by: "G. Gopakumar" [gopakumar.gopinadhan- -gmail.com] --20cf307f3996212c6a04ca73c118 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: quoted-printable Dear CCL members, Nearly a month ago, I posted a message in CCL regarding the mismatch of dihedral angles (gas-phase calculation vs crystal structure). In response to my question, several people replied (including private email= s and conversations). After carefully going through the comments and studying my system in detail (playing around the dihedrals and all), I came to the conclusion that the mismatch (in my case) was due to crystal packing. In order to keep the CCL culture alive, I am posting the summary of responses. Thanks a lot everyone who has replied (including private conversations), and suggested useful references. With sincere thanks and best regards Gopakumar ---------------------------------------------------------------------------= ------------------------------------ Question: =3D=3D=3D=3D=3D=3D=3D Dear CCL members, I was calculating a molecule (71 atoms) for which the crystal structure is known. However, according to my calculation, one of the dihedral angles shows large mismatch (deviation of 12 degrees). I repeated the calculation with both B3LYP and BP86 functional and the values are given below: B3LYP/def2-TZVP ~ 21.2 BP86/def2-TZVP ~ 20 Experimental ~ 8.1 I would not worry too much if it was 2-3 degree deviation, but 12 is something that I should be careful with. I was wondering, whether there is a chance for an experimental error? Or is it just that DFT overestimates the dihedral angle? Does any body know any references where people encountered similar problem. I would be thankful for your comments with kind regards Gopakumar ---------------------------------------------------------------------------= ------------------------------------- Responses: =3D=3D=3D=3D=3D=3D=3D=3D Hi I had something similar and in my case, the difference could be attributed to crystal packing in the solid state whereas my calculations were in the gas phase so there were no intermolecular interactions. Perhaps this applies to you? Good luck Jacquie ---------------------------------------------------------------------------= ------------------------------------- Dear Gopakumar, If you are looking at a 71 atom molecule I would not be surprised if there were some medium-range dispersion interactions in play that could influence the dihedral angle you are curious about. Perhaps try using a DFT flavor with dispersion corrections. Best regards, --john ---------------------------------------------------------------------------= ------------------------------------- Dear Gopakumar, You did not mention but I guess you have studied your molecule in gas phase by DFT (if not please provide some details). You should not expect to see the solid phase geometry in gas phase! In solid phase several effects particularly influence of other molecules in the crystal packing affect the geometry. A famous example in this case is biphenyl. You may know that in the gas phase the dihedral angle between two phenyl rings is ~44 degrees but in solid phase the angle reduces to 0! If you compute geometry of biphenyl at any DFT or any other computational level you will see that the solid state geometry is a first order saddle point, i.e. TS. Be careful in interpretation of gas phase computations and comparing them with X-ray crystallography data... Good luck, Cina ---------------------------------------------------------------------------= ---------------------------------------------- Sent to CCL by: "Eike Huebner" [ccl*eike-huebner.de] Usually, in these cases the energy difference of the structures at the given dihedral angles is very small. If you are interested a bit more in detail, you could calculate the rotational barrier (at fixed dihedral angles, in steps > from something around 5 =B0 beginning at the energy minimum for example) around your dihedral angle. In most of these cases, you will find the energy difference between the minimum and the crystal structure to be in the range of some kJ/mol, which is easily compensated by crystal packing effects, pi-stacking, hydrogen bonding and so on. ---------------------------------------------------------------------------= ----------------------------------------------- I fully agree with Cina. We recently published a study of several hundred instances of small molecules bound to protein structures in the PDB (Sitzmann et al., PDB Ligand Conformational Energies Calculated Quantum-Mechanically, J. Chem. Inf. Model., 52(3), 739-756, 2012). In this context, we also looked at a number of small-molecule crystal structures in the CSD. The upshot was that significant changes both in geometry and energy vs. the vacuum (or gas phase) structure are possible in any crystal environment. Marc ---------------------------------------------------------------------------= ------------------------------------------------ Mark Zottola mzottola*o*gmail.com While in the main I can agree that molecules in a crystal structure can have different geometries than those in the gas phase, there is usually a reason why the difference exists. However, just using the idea of solid state effects as a nebulous rationalization for those differences is not justified nor intellectually honest. TO be clear, I fully agree with with Dr. Nicklaus' statement that "significant changes both in geometry and energy vs. the vacuum (or gas phase) structure are possible in any crystal environment". However, in relation to the original poster, just because differences can occur is not sufficient rationalization for a difference. ---------------------------------------------------------------------------= ------------------------------------------------- Dear Gopakumar, Like Cina said it would be impossible to reproduce the geometry of a molecule in the solid phase with gas phase computations. However from my experience calculations on molecules in solvent might be able to predict the geometry in solid phase. So you can try to do computations on your molecule using DFT with the PCM solvation model and the UFF force field. Adeyinka Adedapo Department of Chemistry, University of Pretoria, South Africa ---------------------------------------------------------------------------= -------------------------------------------------- I would look at the unit cell for the crystal and see how the dihedrals are situated. Often it is obvious why those particular structural features differ significantly from phase to phase (and sometimes it is not so obvious). Cheers, Flick _______________ William F. Coleman Professor of Chemistry Wellesley College Wellesley MA 02481 =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D --20cf307f3996212c6a04ca73c118 Content-Type: text/html; charset=ISO-8859-1 Content-Transfer-Encoding: quoted-printable Dear CCL members,

=A0=A0=A0=A0 Nearly a month ago, I posted a messag= e in CCL regarding the
mismatch of dihedral angles (gas-phase calculatio= n vs crystal structure).
In response to my question, several people rep= lied (including private emails
and conversations). After carefully going through the comments and
stud= ying my system in detail (playing around the dihedrals and all), I came to the conclusion that the mismatch (in my case) was due to
crystal pa= cking.

In order to keep the CCL culture alive, I am posting the summary of responses. Thanks a lot everyone who has replied (including private
conversations), and suggested useful references.

With sincere thank= s and best regards

Gopakumar


----------------------------= ---------------------------------------------------------------------------= --------
Question:
=3D=3D=3D=3D=3D=3D=3D

Dear CCL members,

=A0=A0=A0=A0 I was calculating a molecule (71 = atoms) for which the crystal structure is
known. However, according to m= y calculation, one of the dihedral angles shows
large mismatch (deviation of 12 degrees). I repeated the calculation with b= oth
B3LYP and BP86 functional and the values are given below:

B3LYP/def2= -TZVP=A0=A0 ~ 21.2
BP86/def2-TZVP=A0=A0=A0=A0 ~ 20
Experimental=A0= =A0=A0=A0=A0=A0=A0=A0=A0 ~ 8.1

I would not worry too much if it was = 2-3 degree deviation, but 12 is something
that I should be careful with.

I was wondering, whether there is a c= hance for an experimental error? Or is it
just that DFT overestimates th= e dihedral angle? Does any body know any
references where people encoun= tered similar problem.

I would be thankful for your comments

with kind regards

= Gopakumar

---------------------------------------------------------= -------------------------------------------------------
Responses:
=3D=3D=3D=3D=3D=3D=3D=3D

Hi

I had something similar and in my case, the difference could be
attributed to crystal packing in the solid state whereas my
calculations were in the gas phase so there were no intermolecular
interactions. Perhaps this applies to you?

Good luck

Jacquie

------------------------------------------------------------= ----------------------------------------------------

Dear Gopakumar,=

If you are looking at a 71 atom molecule I would not be surprised if
there were some medium-range dispersion=20 interactions in play that
could influence the dihedral angle you are=20 curious about.=A0 Perhaps
try using a DFT flavor with dispersion=20 corrections.=A0

Best regards,

--john


-------------------------------= ---------------------------------------------------------------------------= ------


Dear Gopakumar,

You did not mention but I guess you have studied your molecule
in gas phas= e by DFT (if not please provide some details). You should
not expect to= =20 see the solid phase geometry in gas phase! In solid
phase several=20 effects particularly influence of other molecules in the
crystal packin= g affect the geometry. A famous example in this case is
biphenyl. You ma= y know that in the gas phase the dihedral angle
between two phenyl rings is ~44 degrees but in solid phase the angle
reduces to 0! If you compute=20 geometry of biphenyl at any DFT or any
other computational level you=20 will see that the solid state geometry is
a first order saddle point,= =20 i.e. TS. Be careful in interpretation of gas
phase computations and com= paring them with X-ray crystallography
data...

Good luck,
Cina

------------------------------------------------= -------------------------------------------------------------------------
Sent to CCL by: "Eike Huebner" [ccl*eike-huebner.de]

Usually, in these cases the energy difference of the structures at the= =20 given
dihedral angles is very small. If you are interested a bit more i= n detail, you
could calculate the rotational barrier (at fixed dihedral= =20 angles, in steps from
something around 5 =B0 beginning at the energy=20 minimum for example) around
your dihedral angle. In most of these cases= , you will find the energy difference
between the minimum and the crysta= l structure to be in the range of some kJ/mol,
which is easily=20 compensated by crystal packing effects, pi-stacking, hydrogen
bonding= =20 and so on.


-----------------------------------------------------= ---------------------------------------------------------------------
I fully agree with Cina.

We recently published a study of several hundred instances of small=20 molecules
bound to protein structures in the PDB (Sitzmann et al., PDB= =20 Ligand
Conformational Energies Calculated Quantum-Mechanically, J. Chem= . Inf. Model.,
52(3), 739-756, 2012). =A0In this context, we also looked= at a number of small-molecule
crystal structures in the CSD.

The upshot was that significant changes both in geometry and energy vs.=20 the
vacuum (or gas phase) structure are possible in any crystal=20 environment.

Marc


-----------------------------------------------------------= ----------------------------------------------------------------

Mark Zottola mzottola*o*gmail.com=A0


=

While in the main I can agree that molecules in a crystal structure can have=20 different
geometries than those in the gas phase, there is usually a=20 reason why the difference
exists. =A0However, just using the idea of so= lid state effects as a nebulous rationalization
for those differences is= =20 not justified nor intellectually honest. =A0

TO be clear, I fully agree with with Dr. Nicklaus'=20 statement that "significant changes
both in geometry and energy vs= . the=20 vacuum (or gas phase) structure are possible in
any crystal=20 environment". =A0However, in relation to the original poster, just bec= ause
differences can occur is not sufficient rationalization for a=20 difference. =A0


--------------------= ---------------------------------------------------------------------------= -----------------------------



<= /div>
Dear Gopakumar,
=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0
=A0=A0=A0=A0=A0= =A0=A0 Like Cina said it would be impossible to=20 reproduce the geometry of a molecule
in the solid phase with gas phase= =20 computations. However from my experience
calculations on molecules in= =20 solvent might be able to predict the geometry in solid phase.
So you ca= n try to do computations on your molecule using DFT with=A0the PCM=20 solvation
model and the UFF force field.
=A0
Adeyinka Adedapo
Department of Chemistry,
University of Pretoria, South Africa

-------------------------------= ---------------------------------------------------------------------------= -------------------


I would look at=20 the unit cell for the crystal and see how the dihedrals are
situated.= =A0=20 Often it is obvious why those particular structural features differ=20
significantly from phase to phase (and sometimes it is not so obvious).=

Cheers,

Flick

_______________
William= F. Coleman
Professor of Chemistry
Wellesl= ey College
Wellesley MA 02481

=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D




--20cf307f3996212c6a04ca73c118-- From owner-chemistry@ccl.net Mon Sep 24 14:33:00 2012 From: "Amandeep Sangha sangha11_._gmail.com" To: CCL Subject: CCL: Internal consistency failure #1 in ROv08 error in Gaussion09 EPT run Message-Id: <-47655-120924131851-24611-TKJJIFISb9WCHI4sikbp0Q+/-server.ccl.net> X-Original-From: "Amandeep Sangha" Date: Mon, 24 Sep 2012 13:18:50 -0400 Sent to CCL by: "Amandeep Sangha" [sangha11]|[gmail.com] Hi there, I want to calculate the ionization potential of a molecule using EPT in G09. When i run the job using following in the route section #P M062X/6-311++G(d,p) EPT I get the following error: Internal consistency failure #1 in ROv08. Does anyone know what is this error and a possible solution for this error? Thanks in advance, amandeep From owner-chemistry@ccl.net Mon Sep 24 15:10:00 2012 From: "Elaine Meng meng%x%cgl.ucsf.edu" To: CCL Subject: CCL: Statistical study of weak bonds in chemical compounds Message-Id: <-47656-120924123020-20083-ONy0kkO/oLyKBxBsU+uX9A{}server.ccl.net> X-Original-From: "Elaine Meng" Date: Mon, 24 Sep 2012 12:30:19 -0400 Sent to CCL by: "Elaine Meng" [meng++cgl.ucsf.edu] Dear Alexandre, Not that they would replace your work (as they are older), but these previous publications may be of interest. I believe both were based on high-resolution subsets of CSD. Three-dimensional hydrogen-bond geometry and probability information from a crystal survey. Mills JE, Dean PM. J Comput Aided Mol Des. 1996 Dec;10(6):607-22. http://www.ncbi.nlm.nih.gov/pubmed/9007693 Tables of bond lengths determined by X-ray and neutron diffraction. Part 1. Bond lengths in organic compounds. Frank H. Allen , Olga Kennard , David G. Watson , Lee Brammer , A. Guy Orpen and Robin Taylor. J. Chem. Soc., Perkin Trans. 2, 1987, S1-S19 http://pubs.rsc.org/en/Content/ArticleLanding/1987/P2/p298700000s1 Elaine ---------- Elaine C. Meng, Ph.D. UCSF Computer Graphics Lab (Chimera team) and Babbitt Lab Department of Pharmaceutical Chemistry University of California, San Francisco